Message boards : ATLAS application : ATLAS native app
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 4 · 5 · 6 · 7

AuthorMessage
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 34990 - Posted: 13 Apr 2018, 15:11:51 UTC

There will probably lots of credits for the first WUs running with the new version, and probably very vew credits for the following batches.
The reason is that the host's fpops counters are reset due to the version change and the current WUs have much longer runtimes than average WUs.
Once this has been adapted and another batch with shorter runtimes will be sent out we will most likely see much less credit per WU.
ID: 34990 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 34992 - Posted: 13 Apr 2018, 15:53:26 UTC

Something strange happened when I tried to run a 2.54 task prior to a couple of 2.53 tasks.

1. Short before a running task was finished I suspended (via GUI) all 2.53 tasks that were uninitialized.
2. As expected, the remaining 2.54 task started as soon as a CPU core became available.
3. When the 2.54 task was up and running I resumed the uninitialized tasks.
4. Although no free CPU core was available, a fresh 2.53 task started and the 2.54 task was set to preempted/suspended.
5. Despite the status in my BOINC client the 2.54 app itself continues running as it would normally do but the BOINC client doesn't show any progress.
ID: 34992 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2090
Credit: 158,890,498
RAC: 125,729
Message 34993 - Posted: 13 Apr 2018, 16:33:15 UTC

Suspend a task in native-App is also a unsolved solution as
the reboot in a VM (for me SL69). It starts from the beginning a second full time.
Is this Software-Change from 2.53 to 2.54 with a reboot of the Linux useful?
ID: 34993 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 34994 - Posted: 13 Apr 2018, 16:46:46 UTC - in response to Message 34993.  

... Is this Software-Change from 2.53 to 2.54 with a reboot of the Linux useful?

You probably want to know if a reboot is required?
Most likely not for machines that were able to run ATLAS native before.
My intention was to run a 2.54 before the weekend to see if it really runs.

The change was done to solve a problem regarding the python version.
Details can be seen here:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4649&postid=34709
ID: 34994 · Report as offensive     Reply Quote
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 34997 - Posted: 13 Apr 2018, 18:52:14 UTC - in response to Message 34994.  

The only change in 2.54 is fixing the python version problem. There should be no need to reboot to run the new version.

But by accident we forgot to mark 2.53 as beta so we corrected that with 2.54. This is why it doesn't show in the applications list because beta apps are not displayed there.
ID: 34997 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 34998 - Posted: 13 Apr 2018, 19:07:06 UTC - in response to Message 34997.  

... by accident we forgot to mark 2.53 as beta so we corrected that with 2.54. This is why it doesn't show in the applications list because beta apps are not displayed there.

That's a pity.
The numbers for 2.53 were very impressive!

Can it be integrated like sixtracktest which is also beta (or, perhaps not really?)?
ID: 34998 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2090
Credit: 158,890,498
RAC: 125,729
Message 35000 - Posted: 13 Apr 2018, 19:53:14 UTC

SL 69 got a SSL secure Python update today.
ID: 35000 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 35019 - Posted: 16 Apr 2018, 7:39:45 UTC - in response to Message 34992.  

Something strange happened ...

Update

My BOINC client worked through the task list step by step in the same order the tasks had been sent by the server.
When it came to the suspended task, the scientific app had already finished it's work and BOINC uploaded/reported it.
Unfortunately due to the lost communication the reported task duration was only a few seconds and therefore the credit reward was very poor.
The stderr.txt shows the correct values:
https://lhcathome.cern.ch/lhcathome/result.php?resultid=187372790

I wonder if anybody from the development team works on that issue as it also affects tasks that are interrupted due to other reasons, e.g. reboots.
ID: 35019 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2090
Credit: 158,890,498
RAC: 125,729
Message 35020 - Posted: 16 Apr 2018, 8:14:26 UTC

Saw this in your task:
THREADS=1
Checking for CVMFS
CVMFS is installed
OS:cat: /etc/redhat-release: Datei oder Verzeichnis nicht gefunden

This is not SLC6, need to run with Singularity....
ID: 35020 · Report as offensive     Reply Quote
computezrmle
Volunteer moderator
Volunteer developer
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 15 Jun 08
Posts: 2401
Credit: 225,518,559
RAC: 122,596
Message 35021 - Posted: 16 Apr 2018, 8:25:34 UTC - in response to Message 35020.  

Saw this in your task:
THREADS=1
Checking for CVMFS
CVMFS is installed
OS:cat: /etc/redhat-release: Datei oder Verzeichnis nicht gefunden

This is not SLC6, need to run with Singularity....

Anything wrong with those messages?
You'll find them in any of my ATLAS native logs.
They just tell you that I'm not running SLC6 and the app will therefore use Singularity.
ID: 35021 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2090
Credit: 158,890,498
RAC: 125,729
Message 35022 - Posted: 16 Apr 2018, 9:08:41 UTC - in response to Message 35021.  

Sorry,
have read you are using SL69 or CentOS!
ID: 35022 · Report as offensive     Reply Quote
maeax

Send message
Joined: 2 May 07
Posts: 2090
Credit: 158,890,498
RAC: 125,729
Message 37600 - Posted: 13 Dec 2018, 13:16:49 UTC - in response to Message 34610.  
Last modified: 13 Dec 2018, 13:18:04 UTC

The slot numbers in Boinc-manager are growing up.
They where all empty (only the folder), but NOT deleted.


Have one Computer with native-App and 32 empty slot-folder.


This is back since the beginning of this month. SL69.
ID: 37600 · Report as offensive     Reply Quote
Previous · 1 . . . 4 · 5 · 6 · 7

Message boards : ATLAS application : ATLAS native app


©2024 CERN